Group-sequential response-adaptive designs for censored survival outcomes

• Published in: Journal of statistical planning and inference Vol. 205; pp. 293 - 305
• Main Authors: Liu, Wenyu, Coad, D. Stephen
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.03.2020
• Subjects: Error-spending function; Information time; Log hazard ratio; Optimal allocation; Power; Weibull response; Log hazard ratio; Weibull response; Information time; Optimal allocation; Power; Error-spending function
• ISSN: 0378-3758; 1873-1171
• DOI: 10.1016/j.jspi.2019.08.004

Previous work on two-treatment comparisons for immediate responses has shown that the use of optimal response-adaptive randomisation with group sequential analysis can allocate more patients to the better-performing treatment while preserving the error rates. In this paper, the application of the combined approach to censored survival responses is investigated and different optimal response-adaptive randomised procedures are compared. For a maximum duration trial, the information level at the final look is usually unpredictable. An approximate information time is defined. Group sequential tests and optimal allocations for two measures of treatment difference are given. Operating characteristics of the combined approach are investigated by simulation, including cases of exponential and Weibull survival responses and redesign of a clinical trial. The results reveal that the existing boundaries for standard group sequential designs derived based on the error-spending approach can be applied as approximate tests to control the overall type I error rate. Compared to the group sequential complete randomisation design, the combined approach is found to retain ethical advantages as in previous work on immediate responses while the power is not adversely affected. •An approximate information time is defined for maximum duration trials.•Both parametric and nonparametric tests are considered.•Standard critical boundaries using the error-spending approach can be applied.•Simulation shows that the overall type I error rate is controlled.•The combined approach reduces the expected numbers of patients and failures.